SIMRACING / CODE MEETS THE COCKPIT
Telemetry, machine learning, and DIY dashboards. Simracing gives me a reason to build—and a very hands-on way to see the result.
01 / THE MILESTONE
MY FIRST AI PROJECTCould I recognize the circuit just from the car’s position? GT7Tracks was my first AI project: a way to turn Gran Turismo 7 telemetry into training data, then use a model to recognize the track while driving.
Python tools capture coordinates into track-labeled CSV files. A TensorFlow model learns from those captures, and a live tool combines predictions over time to select a track.
FROM TELEMETRY TO RECOGNITION
Car positions → labeled track CSVs
Track samples → TensorFlow model
Live positions → predictions over time
02 / THE BUILDS
TELEMETRY / SOFTWARE
A Python UDP proxy that sends a heartbeat to Gran Turismo 7, receives and decrypts the telemetry stream, and forwards it to an XSim plugin.
Packet logging and CSV export also make it useful for replay and analysis. It builds on the telemetry reverse-engineering work shared by the GTPlanet community.
DRIVER FEEDBACK / HARDWARE
A DIY Arduino dashboard for GameDash, combining gear and numeric displays with an RGB RPM LED strip.
Its serial protocol controls display zones, gear, RPM, brightness, and RPM learning. Settings are stored in EEPROM, and module order can be adapted to the physical build.
Explore DashX2 ↗03 / COMMUNITY CONTRIBUTIONS
I also contributed many plugins for console games to SimTools. These contributions sit alongside my own repositories as another part of my simracing work: connecting game telemetry with motion simulation.
Explore SimTools ↗04 / IN THE DRIVER’S SEAT
My YouTube channel includes demos of some of the projects shared on GitHub.
WATCH ON YOUTUBE ↗GRAN TURISMO 7 / SIMTOOLS
A demo of GT7 telemetry with SimTools on the Spa circuit.
WATCH ON YOUTUBE ↗